Randomized clinical trial examining duration of voucher-based reinforcement therapy for cocaine abstinence

Abstract

Background

This is the first study to systematically manipulate duration of voucher-based reinforcement therapy (VBRT) to see if extending the duration increases abstinence during and following VBRT.

Methods

We randomized cocaine-dependent methadone-maintained adults to Standard (12 weeks; n = 62) or Extended (36 weeks; n = 68) VBRT and provided escalating voucher amounts contingent upon urinalysis verification of cocaine abstinence. Urinalysis was scheduled at least every 2 weeks during the 48-week study and more frequently during VBRT (3/week) and 12 weeks of Aftercare (2/week).

Results

Extended VBRT produced longer durations of continuous cocaine abstinence during weeks 1–24 (5.7 vs 2.7 weeks; p = 0.003) and proportionally more abstinence during weeks 24–36 (X² = 4.57, p = .03, OR = 2.18) compared to Standard VBRT. Duration of VBRT did not directly predict after-VBRT abstinence; but longer continuous abstinence during VBRT predicted abstinence during Aftercare (p = 0.001) and during the last 12 weeks of the study (p < 0.001). Extended VBRT averaged higher monthly voucher costs compared to Standard VBRT ($96 vs $43, p < .001); however, the average cost per week of abstinence attained was higher in the Standard group ($8.06 vs $5.88, p < .001). Participants in the Extended group with voucher costs exceeding $25 monthly averaged 20 weeks of continuous abstinence.

Conclusions

Greater abstinence occurred during Extended VBRT, but providing a longer duration was not by itself sufficient to maintain abstinence after VBRT. However, if abstinence can be captured and sustained during VBRT, then providing longer durations may help increase the continuous abstinence that predicts better long-term outcomes.

Introduction

Contingency management (CM) interventions, including voucher-based reinforcement therapy (VBRT), are among the most efficacious methods for improving drug abstinence during drug abuse treatment and have been identified as an empirically-based treatment approach for both opiate and cocaine drug use disorders (Chambless et al., 1998, Chambless and Ollendick, 2001). Meta-analyses of CM yield small to medium overall effect sizes for reducing opiate use during methadone treatment (r = .25; Griffith et al., 2000) and VBRT yields moderate effect sizes for outpatient treatment of cocaine use (r = .35; Lussier et al., 2006).

In VBRT, each time patients provide a urine sample testing negative for specified drugs, they are given a voucher that can be exchanged for a range of goods and services. Typically, the value of the voucher increases gradually with each consecutive drug-free urine sample provided. A drug positive sample or failure to provide a scheduled sample results in the voucher value being reset to the initial value from which it can again escalate according to the same rules. The majority of VBRT studies addressing illicit drug use have implemented a 12-week escalating schedule of voucher delivery.

In surveys of community-based treatment providers, 15–22% of respondents indicate they believe that the effects of CM disappear after the intervention ends (Kirby et al., 2012). This belief is not unfounded; animal research has repeatedly demonstrated that behavior reverses toward baseline after reinforcement is terminated (Skinner, 1938). However, these studies are conducted in operant chambers that minimize extraneous variables, producing an environmental vacuum. In applied research, reinforcement may occur in a social context where individuals are exposed to many factors that can function as naturally-occurring reinforcers (see Baer, 1982) for both drug use and abstinence. In drug abuse treatment research, drug use often returns toward pretreatment levels when CM is abruptly terminated, suggesting that naturally-occurring reinforcers for abstinence are not present (e.g., Silverman et al., 1996, Silverman et al., 1999). However, 12 weeks may be too short for changes in the surrounding environment to occur that will reinforce and maintain abstinence.

Silverman et al. (2004) provided support for the view that delivering a longer CM intervention may result in better maintenance of abstinence. Relative to usual care with or without contingent take-home doses, patients receiving contingent methadone take-home doses and VBRT for a full year did not show precipitous decreases in abstinence during the 8 weeks after VBRT was terminated. Also, Higgins et al. (2000) reported that maximum duration of continuous cocaine abstinence during treatment predicted longer-term cocaine abstinence at 6, 9, and 12 months after treatment entry, suggesting that there is no fixed amount of abstinence needed to increase the odds of longer-term abstinence: the odds continue to increase as a function of the amount of during-treatment abstinence achieved.

Although previous studies have provided VBRT for longer than a12-week duration (e.g., Preston et al., 2001, Silverman et al., 2004), this is the first study to systematically manipulate duration of VBRT in a 48-week randomized controlled trial conducted in a community-based methadone treatment program (also see Carpenedo et al., 2010). The purposes of the study were to compare a standard duration (12 weeks) of VBRT to an extended duration (36 weeks) to test the following hypotheses: (1) Compared to Standard VBRT, Extended VBRT will result in (a) longer durations of continuous cocaine abstinence and (b) increased proportions of cocaine-negative urine samples; (2) Extended VBRT participants will show less cocaine use than Standard VBRT participants during (a) a 12-week Aftercare period following VBRT and (b) during the last 12 weeks of the study (i.e., weeks 37–48); and (3) The longest duration of continuous abstinence achieved during VBRT will predict (a) the percent of cocaine-negative samples provided during Aftercare and (b) during the last 12 study weeks.